The present invention relates to filters that improve the precision of multivariate optical computing Multivariate optical computing (MOC) is a predictive spectroscopy technique that incorporates a multi-wave-length spectral weighting directly into analytical instrumentation. The measurement precision of MOC has been studied for various techniques, several of which involve the use of an interference filter described as a multivariate optical element (MOE). Since MOE-based MOC uses detectors that see all wavelengths simultaneously—including wavelengths that carry no information—measurement noise is reduced and measurement precision is increased if the system can be made sensitive primarily to wavelengths carrying information.
In absorption/transmission/reflection measurements, the best measurement precision occurs when the detector responds only to those wavelengths where the sample analyte exhibits absorption. Thus, the ideal detector response would be one that only accumulates a signal where variance related to the analyte occurs in a data set. Such detection schemes are possible with thermal measurements as compared to purely optoelectronic detectors.